JP3392577B2 - Image reading device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image reading device.

[0002]

2. Description of the Related Art Conventionally, an image reading device such as an image reading unit or an image scanner of a digital copying machine can read and input image data. Some image reading apparatuses that digitally input current image data include a model that distinguishes the image data into multi-valued expression pattern data and binary expression character data, and executes optimum image processing for each.

Such image reading devices are disclosed, for example, in Japanese Patent Laid-Open Nos. 5-176167, 5-308516, and 6-15.
[0059] Japanese Patent Publication No. The image processing apparatuses disclosed in these publications identify image data into pattern data and character data, execute image processing such as dither processing with priority on gradation for the pattern data, and perform resolution on the character data. Image processing such as binarization processing with priority is executed.
In this case, even if pattern data and character data are mixed in the image data, since the pattern data is expressed in gradation and the character data is expressed in binary, high quality image data can be obtained.

Further, there is a model of such an image reading apparatus in which the reading density can be switched. For example,
In the structure in which the line sensor is opposed to the contact glass on which the original is placed via the scanning optical system, the switching of the reading density in the main scanning direction is realized by the switching of the reading process of the line sensor, and the reading density in the sub-scanning direction. Is switched by switching the moving speed of the scanning optical system.

More specifically, the standard reading density is 400 (dp
In the case of (i; dot per inch), if the image data read and input by the line sensor is thinned out in half and the moving speed of the scanning optical system is doubled, the read density becomes half (200 dpi). Similarly, the image data of the line sensor is doubled and complemented,
If the moving speed of the scanning optical system is halved, the reading density will be doubled to 800 (dpi).

In this case, if the user desires, the reading density is increased to read and input the image data precisely, or
It is possible to arbitrarily select to reduce the reading density and read and input the image data at high speed. Further, when the capacity of one image data is fixed, the reciprocal of the reading density as described above corresponds to the scaling factor of the image data. Generally, since an image reading device such as an image scanner is connected to an external device such as a host computer or a laser printer, image data may be scaled by these external devices. The switching of the reading density is effective as described above.

[0007]

In the image reading apparatus as described above, image data in which picture data and character data are mixed can be read and input with good quality, and the reading density of the image data can be switched if desired. be able to.

However, the pattern data and the character data are discriminated from each other by the pattern matching, but the accuracy of the pattern matching may decrease if the reading density changes.

More specifically, when identifying character data, the blank portion is detected as a white background of a constant length by pattern matching. However, if the reading density is lowered, the image data is also reduced relatively. In some cases, a portion that is originally detected as a white background of a certain length becomes smaller than the certain length and cannot be detected. In this case, since the character data is expressed in gradation as the pattern data, the resolution of the character data is lowered and the image quality is lowered.

Furthermore, when identifying character data, a white background area inside the character data is detected by expansion processing of a plurality of lines,
Although the portion in which the density change occurs in the white background area is detected as the edge area of the character data, the contour portion of the pattern data is easily erroneously detected as the edge area of the character data. In this case, the outer peripheral portion of the halftone pattern data is represented in binary and a contour line is generated, but when the reading density is reduced and the image data is relatively reduced, the contour line is relatively enlarged. Image quality is degraded.

The image data read and input by the image reading device may be enlarged by an external device. In such a case, however, the gradation of the characters and the outline of the pattern are enlarged and displayed. , The image quality is remarkably deteriorated.
For example, if the reading density of the image reading device is made to correspond to the printing density or display density of the external device, the deterioration of the image quality does not cause a visual problem, but the reading capacity is limited due to the storage capacity of the image data and the reading speed. It may be necessary to reduce the density.

[0012]

The invention according to claim 1 is
A data input means for reading and inputting image data in which at least one of binary-valued character data and multi-valued pattern data is present is provided, and a white background pixel having a predetermined length with respect to a target pixel sequentially extracted from the image data. The presence or absence of adjacency is determined by pattern matching with a preset mask, and a white background determination unit is provided to detect a target pixel having a predetermined white pixel adjacent to it as a candidate active pixel of the white area. The presence or absence of active pixels in a predetermined peripheral area is determined by expansion processing using a preset mask, and expansion processing means for detecting a target pixel in which the active pixel is located in the predetermined peripheral area as a white background pixel is provided. An edge detection unit for detecting an area is provided, and when both the white background area and the edge area are established, the image data is converted to character data. Character processing for determining the image data as pattern data when one of the white background area and the edge area is not established, and executing image processing in which the resolution is prioritized over the gradation of the determined character data. Means is provided, and pattern processing means for executing image processing in which gradation is prioritized over resolution is provided for the determined pattern data, and density switching means for switching the reading density of the data input means is provided to support switching of the reading density. Then, expansion processing switching means for switching the size of the mask of the expansion processing means is provided.

According to a second aspect of the present invention, there is provided data input means for reading and inputting image data in which at least one of character data represented by binary representation and pattern data represented by multivalued is present.
The presence or absence of adjacency of a white background pixel of a predetermined length to the target pixel sequentially extracted from the image data is determined by pattern matching with a preset mask, and the target pixel adjacent to the white background pixel of the predetermined length is determined as a white background region. A white background determining unit for detecting as a candidate active pixel is provided, and presence / absence of an active pixel in a predetermined surrounding area is determined for the target pixel by expansion processing using a preset mask, and the active pixel is positioned in the predetermined surrounding area. Expansion processing means for detecting the target pixel as a white background pixel is provided, edge detection means for detecting an edge area from the image data is provided, and when both the white background area and the edge area are established, the image data is determined to be character data, and the white background area is determined. An image determination unit that determines the image data as the pattern data when it is not satisfied with one of the edge regions Data is provided with character processing means for executing image processing with priority on resolution over gradation, and pattern processing means for executing image processing with priority on gradation over resolution for the determined pattern data is provided. The density switching means for switching the reading density is provided, and the white background determination switching means for switching the size of the mask of the white background determination means is provided corresponding to the switching of the reading density.

[0014]

According to the first aspect of the invention, when the data input means reads and inputs the image data in which at least one of the binary-represented character data and the multi-valued picture data is present, the white background judgment means extracts from the image data. For the target pixels extracted in order, the presence or absence of adjacency of a white background pixel of a predetermined length is determined by pattern matching with a preset mask, and the target pixel to which the white background pixel of a predetermined length is adjacent is activated as a candidate for the white background area. It is detected as a pixel. Next, the expansion processing means determines whether or not there is an active pixel in a predetermined surrounding area with respect to the target pixel by expansion processing using a preset mask, and selects a target pixel in which the active pixel is located in the predetermined surrounding area as a white background pixel. Therefore, the white background area is detected from the image data.

Since the edge detection means detects the edge area from the image data, the image determination means determines that the image data is character data when both the white background area and the edge area are satisfied, and one of the white background area and the edge area is determined. However, if not established, the image data is determined to be the pattern data. Character processing means
The determined character data is subjected to image processing in which resolution is prioritized over gradation, and the pattern processing means performs image processing in which the determined pattern data is prioritized for gradation over resolution. Optimal image processing is executed for each data.

When the density switching means switches the reading density of the data input means, the expansion processing switching means switches the size of the mask of the expansion processing means in response to the switching of the reading density. It is possible to prevent erroneous detection as an edge area of data.

According to a second aspect of the present invention, when the data input means reads and inputs the image data in which at least one of the binary-valued character data and the multi-valued pattern data exists, the white background judgment means extracts the image data from the image data. For the target pixels extracted in order, the presence or absence of adjacency of a white background pixel of a predetermined length is determined by pattern matching with a preset mask, and the target pixel to which the white background pixel of a predetermined length is adjacent is activated as a candidate for the white background area. It is detected as a pixel. Next, the expansion processing means determines whether or not there is an active pixel in a predetermined surrounding area with respect to the target pixel by expansion processing using a preset mask, and selects a target pixel in which the active pixel is located in the predetermined surrounding area as a white background pixel. Therefore, the white background area is detected from the image data.

Since the edge detection means detects the edge area from the image data, the image determination means determines that the image data is character data when both the white background area and the edge area are satisfied, and one of the white background area and the edge area is determined. However, if not established, the image data is determined to be the pattern data. Character processing means
The determined character data is subjected to image processing in which resolution is prioritized over gradation, and the pattern processing means performs image processing in which the determined pattern data is prioritized for gradation over resolution. Optimal image processing is executed for each data.

When the density switching means switches the reading density of the data input means, the white background determination switching means switches the mask size of the white background determination means in response to the switching of the reading density, so that the density change occurs in the white background area. When the above-mentioned portion is detected as the edge area of the character data, even if the contour portion of the pattern data is erroneously detected as the edge area of the character data, the contour line of the pattern generated by the binary representation of the outer peripheral portion is not enlarged.

[0020]

An embodiment of the present invention will be described below with reference to the drawings. First, the image scanner 1 which is the image reading apparatus of the present embodiment mechanically has a contact glass 3 on the upper surface of the main body housing 2 as shown in FIG. A document pressure plate 4 is arranged so as to be openable and closable. A scanning optical system 5 is disposed at a position facing the contact glass 3 from below, and the scanning optical system 5 is slender in the main scanning direction and moves in the sub scanning direction at a ratio of 2: 1. It consists of two running bodies 6 and 7. The first traveling body 6 includes a pair of straight tube fluorescent lamps 8 and one reflecting mirror 9, and the second traveling body 7 includes a pair of reflecting mirrors 10 and 11. The second traveling body 7
Since the line sensor 13 is disposed at a position facing in the sub-scanning direction via the imaging optical system 12, the scanning optical path communicating from the contact glass 3 to the line sensor 13 has the traveling bodies 6 and 7. It is formed by the reflection mirrors 9 to 11. The line sensor 13 has a large number of CCDs (Charge Coupled Devices) as light receiving elements, and these CCDs are connected in the main scanning direction.

An opening 14 is formed outside one end of the contact glass 3 in the sub-scanning direction.
A document conveying mechanism 15 is detachably mounted at a position facing from above. The document transport mechanism 15 has a paper feed tray 16 and a paper discharge tray 17, and the paper discharge tray 17 is integrated with the upper surface of the document pressure plate 4.
A large-diameter platen roller 18 is arranged at a position where the paper feed tray 16 communicates with the paper discharge tray 17, and the platen roller 18 is located above the opening 14. The paper feed tray 16 and the platen roller 18
A separation / feeding mechanism 19 is provided in the gap between and, and a large number of feed rollers 20 are arranged around the platen roller 18.

Then, the image scanner 1 of the present embodiment.
As the image reading mode of a document (not shown), the document fixing mode and the document conveying mode can be switched. In the document fixing mode, the document is fixedly positioned on the contact glass 3 to move the traveling bodies 6 and 7 in the sub-scanning direction, and in the document conveying mode, the traveling bodies 6 and 7 are located below the opening 14. The original is conveyed in the sub-scanning direction by the original conveying mechanism 15 while being fixedly positioned at. Since the image scanner 1 of the present embodiment scans and scans the image data of the original in the main scanning direction and the sub scanning direction as described above, a data input unit for reading and inputting the image data is realized.

A drive motor 21 for driving the scanning optical system 5 and the like is disposed at one end of the main body housing 2 in the sub-scanning direction, and various data processing is performed on the bottom of the main body housing 2. A circuit board 22 to be executed is arranged.

The image scanner 1 having the above structure.
1, is connected to an external device 23 such as a host computer and a laser printer, and forms an image processing system 24 as a whole. An image processing circuit 25 is connected to the line sensor 13, and the image processing circuit 25 is connected to the external device 23. The image processing unit 25 identifies the image data read and scanned from the original by the line sensor 13 into binary representation character data and multi-value representation pattern data, and the optimum image processing for each of these data. To execute.

Therefore, the image processing circuit 25 has a separation processing unit 26, a binarization processing unit 27, an intermediate processing unit 28, and a gradation processing unit 29, and these processing units 26 to 29.
Are connected to the selection circuit 30. As will be described later in detail, the separation processing unit 26 divides the image data into a resolution priority area corresponding to character data, a gradation priority area corresponding to picture data, and an intermediate processing area not classified into these. Identify. The binarization processing unit 27 executes the image processing in which the resolution is prioritized over the gradation in the image data, and the intermediate processing unit 28 performs the image processing in which the resolution and the gradation are compatible with the image data. Then, the gradation processing unit 29 executes image processing in which gradation is prioritized over resolution in the image data.

The selection circuit 30 includes the separation processing unit 26.
Since the image data in the resolution priority area is selectively input from the binarization processing unit 27 in accordance with the identification result of, the character processing means for executing the image processing in which the resolution is prioritized over the tone is applied to the character data. Has been realized. Similarly, the selection circuit 30 inputs the image data of the intermediate processing region from the intermediate processing unit 28 and the image data of the gradation priority region from the gradation processing unit 29. In addition, a pattern processing means for executing image processing in which gradation is prioritized over resolution is realized.

As described in detail in FIG. 3, the separation processing section 26 for identifying the image data into various areas as described above is an edge area detecting means which is an edge detecting means connected in parallel to the line sensor 13. It has a circuit 31 and a white background area detection circuit 32, and these area detection circuits 31, 3
2 is connected to one area determination circuit 33 which is an image determination means. The edge area detection circuit 31 detects a portion where white pixels are continuous and a portion where black pixels are continuous from the image data, and detects a position where these portions are present as an edge area. Note that such an edge detection circuit 31
The processing operation described in JP-A-5-1761 is known.
It is disclosed in detail in Japanese Patent Publication No. 67 and the like.

As will be described later in detail, the white background area detection circuit 32 determines the presence or absence of a white background of a predetermined length adjacent to the pixel of interest, and detects the direction if it exists. The area determination circuit 33 outputs, for each pixel of the image data, an identification signal of the resolution priority area, the intermediate processing area, and the gradation priority area based on the combination of the output results of the detection circuits 31 and 32.

As described above, the white background area detection circuit 32 for detecting a white background, as shown in FIG.
n Transfer Function) correction circuit 34, binarization circuit 35,
It has a white background pixel detection circuit 36 which is a white background determination means, an expansion processing circuit 37 which is an expansion processing means, and a correction processing circuit 38. As shown in FIG. 7, the MTF correction circuit 34 has an MTF filter of a predetermined shape set in advance,
The image data is sharpened by the correction calculation of the MTF filter. The binarization circuit 35 is preset with a predetermined threshold value, and compares the threshold value with the threshold value to binarize the image data into black and white. As shown in FIG. 8, the white background pixel detection circuit 36 is preset with a mask such as “1 × 5” or “5 × 1”, and a pattern of a predetermined length is obtained by pattern matching between the mask and the image data. A pixel of interest adjacent to a white background is detected as an active pixel.

The structure and processing operation of the white background pixel detection circuit 36 will be described in detail below with reference to FIG. This white background pixel detection circuit 36 has twenty-five FF (Flip Flop) circuits 39 corresponding to a “5 × 5” mask, and each row of these FF circuits 39 has a predetermined capacity.
They are connected in order through one by one. The FF circuit 39 is connected to a detection operation circuit 47 including two AND gates 41 to 44 of two systems and two OR gates 45 and 46 of two systems. The output is connected to one selector circuit 48.

The first AND gate 41 of the first system
Is connected to the FF circuit 39 at a position forming a “1 × 5” mask, and the second AND gate 42 of the first system is connected to the FF circuit 39 by “5 × 5”.
The two AND gates 43 and 44 of the second system are connected to a position for forming a 1 "mask.
It is connected to the FF circuit 39 at a position where a mask of “1 × 3” and “3 × 1” is formed. These two-system AND gates 41 to 44 are individually connected to the two-system OR gates 45 and 46, and the control inputs of the selector circuit 48 to which these OR gates 45 and 46 are connected are not shown. CPU (Central Processing
Unit) is connected.

As will be described later in detail, this CPU is
Each unit is controlled by various data processing. For example, by switching the reading processing of the line sensor 13, the scanning speed of the scanning optical system 5, the transportation speed of the document transportation mechanism 15, etc. Read density of 4
It can be switched to two levels of 00 and 200 (dpi). Here, the density switching means is realized, and the CPU sends the read density switching signal as described above to the selector circuit 48.
Output to.

The image data sequentially input to the white background pixel detection circuit 36 for each pixel is processed by the FF circuit 39.
It is temporarily stored so that the pixel of interest is located at the center of “5 × 5”. At this time, the AND gates 41 and 4 of the first system
2 detects "1 × 5" and "5 × 1" white background pixels centered on the pixel of interest based on the logical product of the image data. Therefore, the OR gate 45 of the first system calculates the logical sum of the detection results. The center of white background pixels that are continuous with “1 × 5” or “5 × 1” is output as an active pixel.

At the same time, the AND gate 4 of the second system
3, 44 are "1x3" and "3x" with the pixel of interest at the center.
Since the white background pixel of 1 "is detected, the OR gate 46 of the second system outputs the center of the white background pixels continuous to" 1x3 "or" 3x1 "as an active pixel. Output to the selector circuit 48. As will be described later in detail, this selector circuit 48
Is one of the two input active pixels,
Since the output is selectively output according to the read density switching signal input from the PU, a white background determination switching unit that switches the size of the mask of the white pixel detection circuit 36 corresponding to the switching of the read density is realized here. .

In the expansion processing circuit 37, as shown in FIG. 9, a mask such as “3 × 9” is set in advance for expansion processing, and even one active pixel is included in the mask centered on the pixel of interest. Is present, the target pixel is detected as a white background pixel. This expansion process is executed to detect a white background pixel inside the character to determine a white background portion where the character is located as a pattern portion.

Here, the structure and processing operation of the expansion processing circuit 37 will be described in detail below with reference to FIG. The expansion processing circuit 37 has twenty-seven FF circuits 49 corresponding to a “3 × 9” mask, and each row of these FF circuits 49 passes through one line memory 50 of a predetermined capacity. Connected in order. The FF circuit 49 is connected to an expansion arithmetic circuit 55 composed of four OR gates 51 to 54, and the expansion arithmetic circuit 55 and one OR gate 52 are connected to one selector circuit 56. The three OR gates 51 to 53 are connected to the FF circuit 49 for each “3 × 9” row, and these OR gates 51 to 53 are individually connected to the OR gate 54. The control input of the selector circuit 56 to which the OR gate 54 and the OR gate 52 are connected,
The CPU is connected.

The image data sequentially input to the expansion processing circuit 37 for each pixel is "3" by the FF circuit 49.
It is temporarily stored so that the pixel of interest is located at the center of × 9 ″. At this time, the OR gates 51 to 53 are “3 × 9”.
By the logical sum of the image data temporarily stored in the FF circuit 49, the expansion operation of “1 × 9” is executed for each row of the three rows of image data centered on the pixel of interest. Then, the OR gate 54 executes expansion operation of “3 × 9” image data centered on the pixel of interest by the logical sum of the output results of the OR gates 51 to 53.

At this time, the OR gate 54 outputs a logical sum of "3.times.9" image data centered on the pixel of interest to the selector circuit 56 and "1.times.9" centered on the pixel of interest. The logical sum of the image data is output from the OR gate 52. Then, the selector circuit 56
Selectively outputs one of the two input calculation results in accordance with the read density switching signal input from the CPU. Therefore, the mask of the expansion processing circuit 37 corresponds to the read density switching here. The expansion processing switching means for switching the size of is realized.

As shown in FIG. 10, in the correction processing circuit 38, a one-dimensional mask is set in advance for the correction processing, and the presence / absence of a white background pixel around the target pixel is conditionally corrected. . That is, when there is a white background on both sides of the pixel of interest due to the pattern matching with the mask described above, (logical sum of a to p) and (logical sum of A to P) = 1, so in this case On the other hand, a 2-bit output signal "10" corresponding to the identification signal of the resolution priority area is output. When a white background exists on one side of the pixel of interest, (logical sum of a to p) or (logical sum of A to P) = 1, and in this case, the output signal “corresponding to the identification signal of the intermediate processing area” 01 "is output, and if not, an output signal" 00 "corresponding to the identification signal of the gradation priority area is output.

The white background area detection circuit 32 including the above-mentioned various circuits 34 to 38 determines the presence or absence of a white background of a predetermined length adjacent to the target pixel together with the direction, generates a 2-bit output signal, and detects the edge area. The circuit 31 detects the edge area from the image data and generates a 1-bit output signal.
3 outputs identification signals of various areas for each pixel of the image data.

That is, the edge detection circuit 31 determines the pixel of interest as an edge region and generates a 1-bit output signal "1", and the white background region detection circuit 32 detects and outputs a white background on both sides of the target pixel. When the signal "10" is generated, the area determination circuit 33 determines the pixel of interest as the resolution priority area and generates a 2-bit identification signal "10". When the edge detection circuit 31 generates an output signal “1” and the white background area detection circuit 32 detects a white background on one side of the target pixel and generates an output signal “01”, the area determination circuit 33 The target pixel is determined to be the intermediate processing region and the identification signal “01” is generated. In the case of combinations other than these, the region determination circuit 33 determines the target pixel to be the gradation priority region and the identification signal “01”. Generates 00 ".

It should be noted that the binarization processing unit 27, which executes the image processing of resolution priority in the resolution priority area of the image data,
As shown in FIG. 1, the MTF correction circuit 57 and the binarization circuit 5
8 and. Like the MTF correction circuit 34 of the white background area detection circuit 32 of the separation processing unit 26, the MTF correction circuit 57 has a predetermined shape of MT as shown in FIG.
The F filter is set, and the image data is sharpened by the correction calculation of the MTF filter. Similarly to the binarizing circuit 35, the binarizing circuit 58 has a predetermined threshold value set in advance, and binarizes the image data into black and white by comparison with this threshold value.

The intermediate processing section 2 for executing image processing in which the resolution and gradation are taken into consideration in the intermediate processing area of the image data.
As shown in FIG. 1, the reference numeral 8 has an MTF correction circuit 59 and a dither processing circuit 60. The MTF correction circuit 59
In the same manner as the MTF correction circuits 34 and 57, the image data is sharpened by the correction calculation of the MTF filter. The dither processing circuit 60, as shown in FIG.
A matrix of x4 "dither processing is set, and the image data is expressed in gradation by the dither processing of this matrix.

As shown in FIG. 1, the gradation processing unit 29 for executing gradation-priority image processing in the gradation-priority area of image data has a smoothing circuit 61 and a dither processing circuit 62. ing. As shown in FIG. 12, the smoothing circuit 61 is preset with a “3 × 5” smoothing filter, and smoothes the image data by the processing of this filter. Then, the dither processing circuit 62, similarly to the dither processing circuit 60, expresses the gradation of the image data by the matrix dither processing.

With such a configuration, the image scanner 1 of the present embodiment reads and inputs the image data of the original document and outputs it to the external device 23 such as a host computer or a laser printer.
Can be output to. At this time, the image scanner 1 of the present embodiment sets the reading density of the image data to 400, 200 (dpi).
The read density is 40
Read and input image data precisely as 0 (dpi),
The reading density can be set to 200 (dpi) to reduce the amount of image data and improve the reading speed. Further, the image scanner 1 of the present embodiment identifies the read and input image data as character data or picture data, and executes optimum image processing for each.

More specifically, in the image scanner 1 of the present embodiment, the image data read and input is separated by the separation processing unit 26.
Is determined as a resolution priority area, an intermediate processing area, and a gradation priority area. The binarization processing unit 27 executes the image processing in which the resolution is prioritized in the resolution priority area, and the intermediate processing unit 28 executes the image processing in which the resolution and the gradation are compatible with each other in the intermediate processing area. In the tonality priority area, the gradation processing unit 29 executes image processing in which gradation is prioritized. Therefore, the character data portion is expressed in binary as a resolution priority area, and the pattern data portion is expressed in gradation as a gradation priority area. Therefore, the read and input image data is optimally image-processed and external device 23 is output. In addition, since the image processing in which both the resolution and the gradation are compatible is executed in the intermediate processing area which is not determined as the character data or the pattern data, the image data is processed extremely well.

When the separation processing section 26 identifies the image data into various areas as described above, the edge area detection circuit 31 is used.
Detects the edge area from the image data, and the white background area detection circuit 32 determines the presence or absence of a white background adjacent to the pixel of interest of a predetermined length. The identification signal of each area is output for each.

When the white background area detection circuit 32 detects a continuous white background as described above, the MTF correction circuit 34
The image data is sharpened by the correction calculation of the preset MTF filter, and the binarization circuit 35 binarizes the image data by the preset threshold value. Next, the white background pixel detection circuit 36 detects a target pixel adjacent to a white background of a predetermined length as an active pixel by pattern matching between a preset mask and image data, so that the expansion processing circuit 37 is set in advance. By the expansion process using the mask, if at least one active pixel exists in the mask centered on the target pixel, the target pixel is detected as a white background pixel. Then, the correction processing circuit 38 outputs a candidate identification signal of each region when a white background portion exists on both sides or one side of the pixel of interest by pattern matching between a preset one-dimensional mask and image data.

As described above, when the white background pixel detection circuit 36 detects a target pixel adjacent to a white background of a predetermined length as an active pixel, the target pixel is positioned at the center of the “5 × 5” FF circuit 39. The image data is temporarily stored in, and by the logical operation of the AND gates 41 and 42 of the first system and the OR gate 45, the pixel of interest at the center where the white background pixels are continuous in the “1 × 5” or “5 × 1” mask It is detected as an active pixel. At this time, the second system AND gate 43,
“1 × 3” by logical operation of 44 and OR gate 46
A pixel of interest at the center of which a white background pixel is continuous with a “3 × 1” mask is also detected as an active pixel.

One of these active pixels is selectively output by the selector circuit 48. This selector circuit 48 has a reading density of 400 (dpi) of the image scanner 1.
In the case of, the active pixels detected by the mask of “1 × 5” and “5 × 1” are selected, and when the reading density is 200 (dpi), the mask of “1 × 3” and “3 × 1” is used. Select the detected active pixel.

Further, as described above, the expansion processing circuit 37 is
When the target pixel is detected as a white background pixel by the expansion processing, the image data is temporarily stored so that the target pixel is located at the center of the “3 × 9” FF circuit 49, and the OR gate 51 is used.
54 to 54, the expansion operation of the image data by the mask of “3 × 9” is executed, and the OR gate 52 executes the expansion operation of the image data by the mask of “1 × 9”.

One of the white background pixels detected by these calculations is selectively output by the selector circuit 56. When the reading density of the image scanner 1 is 400 (dpi), the selector circuit 56 outputs " The white background pixels detected by the 3 × 9 ″ mask are selected, and when the reading density is 200 (dpi), the white background pixels detected by the “1 × 9” mask are selected.

As described above, in the image scanner 1 of this embodiment, the masks of the white background pixel detection circuit 36 and the expansion processing circuit 37 are switched according to the read density of the image data. Even when the density is lower than the reproduction density of the external device 23, it is possible to prevent the quality of the image data from deteriorating. This will be described in detail below with reference to FIGS. 13 and 14.

First, when the white background pixel detection circuit 36 detects a target pixel adjacent to a white background of a predetermined length as an active pixel, as shown in FIG. , “1 × 5” mask and pattern matching. At this time, the read density is 40
If the image data is reduced from 0 (dpi) to 200 (dpi), the image data is relatively reduced. Therefore, as shown in FIG. 13B, the portion where the white background pixels of the character data are continuous is “1 × 5”. It cannot be detected by the mask of ".

However, in the image scanner 1 of the present embodiment, when the reading density is switched to 200 (dpi) as described above, the pattern matching mask is also switched to "1 × 3", so that the white background pixel of the character data is changed. The part where is continuous is detected well. Therefore, the white background portion of the character data is not detected, and the character data is prevented from being erroneously detected by the pattern data and expressed in gradation, so that the deterioration of the image quality due to the decrease of the resolution of the character data is prevented.

Further, the expansion processing circuit 37 detects the target pixel as a white background pixel by the expansion processing to detect the edge area of the character data. However, as shown in FIG. When it appears suddenly, the outline of this pattern is erroneously determined as the edge area of the character data. Even the image scanner 1 of the present embodiment cannot prevent such an erroneous determination, but it is possible to prevent the contour line from expanding on the outer peripheral portion of the pattern data due to this erroneous determination.

That is, the reading density of the image scanner 1 of this embodiment and the reproduction density of the external device 23 are both 400 (dp).
In the case of i), as shown in FIG. 14B, since the portion of the sixth main scanning line is determined to be a white background area, the seventh and eighth areas are expanded by the expansion process using the “3 × 9” mask. The portion of the main scanning line of is also determined as the white background area. At this time, density changes occur in these main scanning line portions, so this is determined as an edge region and binary processing is performed, so a black contour line is formed in the pattern data. However, this contour line is 400 (dp
Since the two main scanning lines of i) are reproduced at 400 (dpi), the deterioration of the image quality is not a problem visually.

On the other hand, the reading density of the image scanner 1 is 20
When the reproduction density of the external device 23 is 0 (dpi) and 400 (dpi), the portion of the third main scanning line is determined to be a white background area as shown in FIG. If the expansion process is performed using a mask of × 9 ″, contour lines are formed at the portions of the fourth and fifth main scanning lines. In this case, this contour is 20
Since it is two main scanning lines of 0 (dpi), when reproduced at 400 (dpi), it becomes four main scanning lines, and the deterioration of the image quality is remarkable.

However, the image scanner 1 of this embodiment
When the reading density is switched to 200 (dpi), the expansion processing mask is also switched to “1 × 9”, so that the expansion processing is not executed in the sub-scanning direction. In this case, since the contour line does not occur in the picture data, deterioration of the image quality is prevented.

In the image scanner 1 of this embodiment, since the mask for expansion processing is switched from "3x9" to "1x9", the expansion processing in the sub-scanning direction is completely stopped and the contour of the pattern data is cut. Although lines are prevented, this may reduce the resolution of the edge area of the character data. For example, if the mask for the expansion process is set to be switched from "5x9" to "3x9", the contour lines generated in the pattern data are reduced from five lines to three lines, so that the image quality This can contribute to the improvement, and the resolution of the edge area of the character data can be maintained well.

In the image scanner 1 of this embodiment, as described above, the white background pixel detection circuit 36, which is the white background determination means, has a mask developed in both the main scanning direction and the sub scanning direction, and at least Although it has been illustrated that the white background of a predetermined length is continuous in one direction, the present invention is not limited to the above embodiment, and a mask developed only in one direction can be set. Is. Also,
Although it has been illustrated that such a mask is shortened in both the main scanning direction and the sub scanning direction according to the reading density, the shortening direction can be set to only one of the main scanning direction and the sub scanning direction. Is.

Similarly, in the image scanner 1 of this embodiment, the expansion processing circuit 37, which is the expansion processing means, has a mask developed in both the main scanning direction and the sub-scanning direction. The invention is not limited to the above embodiment, and it is possible to set a mask developed in only one direction. Further, although it has been illustrated that such a mask is shortened only in the sub-scanning direction corresponding to the reading density, this shortening direction is set as the main scanning direction or both the main scanning direction and the sub-scanning direction. It is also possible.

Further, in the image scanner 1 of this embodiment, when the expansion processing circuit 37, which is expansion processing means, shortens the mask in the sub-scanning direction, the expansion processing is sub-processed by setting the range in the sub-scanning direction as one line. It has been illustrated that it is not executed in the scan direction. In this case, since the contour line is not formed in the pattern data, the image quality of the pattern data is improved, but the edge area of the character data may be expressed in gradation. That is, the degree of mask shortening as described above is optimally set according to various conditions, and for example, the expansion process of five lines can be shortened to three lines.

[0064]

According to the first aspect of the present invention, the presence or absence of active pixels in a predetermined peripheral area of the target pixel is determined by expansion processing using a preset mask, and the active pixel is positioned in the predetermined peripheral area. The expansion processing means for detecting the target pixel as a white background pixel is provided, and the expansion processing switching means for switching the mask size of the expansion processing means in response to the switching of the reading density is provided. Since the generated contour line can be reduced according to the reading density, it is possible to prevent the quality of the image data from being deteriorated even when the reading density of the image data is switched to a value smaller than the reproduction density.

According to a second aspect of the present invention, the presence or absence of adjacency of a white background pixel having a predetermined length with respect to the target pixel sequentially extracted from the image data is determined by pattern matching with a preset mask, and the predetermined length is determined. By providing the white background determination means for detecting the target pixel adjacent to the white background pixel as an active pixel of a candidate for the white background area, and providing the white background determination switching means for switching the mask size of the white background determination means in response to the switching of the read density. Since it is possible to prevent the detection accuracy of the white background region from being lowered due to the switching of the reading density, it is possible to prevent the quality of the image data from being deteriorated even when the reading density of the image data is switched to be smaller than the reproduction density.

[Brief description of drawings]

FIG. 1 is a block diagram showing an image processing system in which an external device is connected to an image scanner which is an image reading device according to an embodiment of the present invention.

FIG. 2 is a vertical cross-sectional side view showing the internal structure of the image scanner.

FIG. 3 is a block diagram showing a separation processing unit.

FIG. 4 is a block diagram showing an area determination circuit.

FIG. 5 is a block diagram showing a white background pixel detection circuit which is white background determination switching means.

FIG. 6 is a block diagram showing an expansion processing circuit that is expansion processing means.

FIG. 7 is a schematic diagram showing an MTF filter set in an MTF correction circuit.

FIG. 8 is a schematic diagram showing a mask set in a white background pixel detection circuit.

FIG. 9 is a schematic diagram showing a mask set in an expansion processing circuit.

FIG. 10 is a schematic diagram showing a mask set in a correction circuit.

FIG. 11 is a schematic diagram showing a matrix set in a dither processing circuit.

FIG. 12 is a schematic diagram showing a filter set in a smoothing circuit.

13A and 13B are schematic diagrams for explaining the pattern matching of the white background pixel detection circuit. FIG. 13A is a read result with a read density of 400 (dpi), and FIG. 13B is a read result with a read density of 200 (dpi). .

FIG. 14 is a schematic diagram illustrating expansion processing of an expansion processing circuit, where (a) erroneously determines the upper edge of the pattern data as the edge area of the character data, and (b) shows a read density of 400 (dpi). )
(C) is a reading result of a reading density of 200 (dpi).

[Explanation of symbols]

32 edge detection means 33 image determination means 36 White background judgment means 37 Expansion processing means

─────────────────────────────────────────────────── ─── Continuation of the front page (58) Fields investigated (Int.Cl. ⁷ , DB name) H04N 1/38-1/409

Claims (2)

(57) [Claims] 1. Data input means for reading and inputting image data in which at least one of binary-valued character data and multi-valued pattern data is present, and a predetermined value for a target pixel sequentially extracted from the image data. White background determining means for determining whether or not a long white background pixel is adjacent by pattern matching with a preset mask, and detecting a target pixel adjacent to a predetermined white background pixel as a candidate active pixel in a white background area; On the other hand, expansion processing means for determining the presence or absence of active pixels in a predetermined surrounding area by expansion processing using a preset mask, and detecting a target pixel in which the active pixels are located in the predetermined surrounding area as a white background pixel, and image data. Edge detection means for detecting an edge area from the image data, and when both the white background area and the edge area are established, the image data is converted to character data. Image determining means for determining image data as pattern data unless one of the white background area and the edge area is satisfied, and a character processing means for executing image processing in which the resolution is prioritized over the gradation of the determined character data. A pattern processing unit that executes image processing in which the gradation is prioritized over the resolution of the determined pattern data; a density switching unit that switches the reading density of the data input unit; and the expansion corresponding to the switching of the reading density. An image reading apparatus comprising: an expansion process switching unit that switches a size of a mask of the processing unit. 2. Data input means for reading and inputting image data in which at least one of binary-valued character data and multi-valued pattern data is present, and a predetermined value for a pixel of interest sequentially extracted from the image data. White background determining means for determining whether or not a long white background pixel is adjacent by pattern matching with a preset mask, and detecting a target pixel adjacent to a predetermined white background pixel as a candidate active pixel in a white background area; On the other hand, expansion processing means for determining the presence or absence of active pixels in a predetermined surrounding area by expansion processing using a preset mask, and detecting a target pixel in which the active pixels are located in the predetermined surrounding area as a white background pixel, and image data. Edge detection means for detecting an edge area from the image data, and when both the white background area and the edge area are established, the image data is converted to character data. Image determining means for determining image data as pattern data unless one of the white background area and the edge area is satisfied, and a character processing means for executing image processing in which the resolution is prioritized over the gradation of the determined character data. A pattern processing unit for executing image processing in which gradation is prioritized over resolution for the determined pattern data, a density switching unit for switching the reading density of the data input unit, and a white background corresponding to the switching of the reading density. An image reading apparatus comprising: a white background determination switching unit that switches a mask size of the determination unit.